My sympathies to those dealing with bytecode changes: Generating the interpreter · Issue #98831 · python/cpython · GitHub
I like what I am seeing. Having an interpreter generator will probably mean the bytecode spec will be clearer, and avoiding error in documentation and inconsistencies in dis module.
if you are positive about it then it’s good sign.
I was wondering whether this would lead to an explosion of new operations, and more work for numba after each CPython release.
If the CPython interpreter can be generated from the bytecode spec, perhaps the Numba Bytecode “interpreter” could also be generated from it. In practice I think this would probably turn out to be hard, but I think it’s worth exploring at an appropriate time.